Harald Hoyer f62e8ac470 perf(llama-cpp-rocm): tune for Strix Halo (gfx1151) ... - Restrict rocmGpuTargets to gfx1151 (Radeon 8060S, RDNA 3.5) — smaller closure, faster compile, no wasted device kernels. - Enable GGML_HIP_ROCWMMA_FATTN: rocWMMA-backed flash attention is a major win on RDNA3+ for the GPU-offloaded attention path. - Enable GGML_HIP_GRAPHS to lower per-token launch overhead. - Add rocwmma to buildInputs to satisfy the WMMA path. llama-server on halo runs with -ngl 99 --flash-attn on, so these flags target the hot path. CPU-side AVX-512 was skipped intentionally — Zen 5 has it, but with full GPU offload the CPU paths barely run.		2026-05-06 09:13:54 +02:00
..
claude-code	chore: claude-code update	2026-05-01 08:15:53 +02:00
gemini-cli	chore: update packages and dependencies	2025-06-30 09:49:01 +02:00
aider-chat.nix	nix fmt	2026-02-24 13:25:42 +01:00
default.nix	perf(llama-cpp-rocm): tune for Strix Halo (gfx1151)	2026-05-06 09:13:54 +02:00
gnome-remote-desktop-mac.patch	add amd	2026-01-17 14:48:45 +01:00
goose.nix	nix fmt	2026-02-24 13:25:42 +01:00
roo-code.nix	nix fmt	2026-02-24 13:25:42 +01:00